AI and the philosophy of mind - know your Claude

​Our AI safety group is meeting on February 18th. Join us to discuss AI safety with fellow EAs. Great for learning more about AI and discussing how to deal with this critical issue.

Bring your laptops!

Know your Claude. We will compile a list of interesting prompts, some of which probe philosophical themes, some of which probe technical themes, and present them to the group one at a time. We will discuss what we think Claude will say in response, and then see what Claude actually says. Then we will reflect a bit on what we think about each prompt, what it might show, how we think Claude should answer, etc. By predicting and then observing how Claude responds, and noticing where our predictions fail, we confront the same epistemic gap that animates the philosophy of mind: the distance between behavioral evidence and the question of what, if anything, it is like to be the system producing that behavior.

​

​URL for the task: https://docs.google.com/document/d/1g-4oRzh3s0-USMTdEvxvm5zcyOQ8bskPoLyXLmVdah4/edit?usp=sharing